Water Striders:The Biomechanics of Water Locomotion and Functional Morphology of the Hydrophobic Sur

Available online at www.sciencedirect.comScienceDirectJoumal of Bionic Engineering 5 (2008) 121-126Water Striders: The Biomechanics of Water Locomotion and FunctionalMorphology of the Hydrophobic Surface (Insecta: Hemiptera-Heteroptera)P. J. Perez Goodwyn', Jin-tong Wang', Zhou-ji Wang, Ai-hong Ji, Zhen-dong Dai', K. Fujisaki?1. Institute for Bio-Inspired Structure and Surface Engineering (IBSS), Nanjing University ofAeronautics and Astronautics,Nanjing 210019, P. R. China2. Laboratory of Insect Ecology, Graduate School of Agriculure, Kyoto University, Kyoto 606-8502, JapanAbstractWater striders are insects living on the water surface, over which they can move very quickly and rarely get wetted. Wemeasured the force of free walking in water stiders, using a hair atached to their backs and a 3D strain gauge. The error wascalculated by comparing force and data derived from geometry and was estimated as 13%. Females on average were stronger(1.32 mN) than males (0.87 mN), however, the ratio of force to weight was not significantly dfferent. Compared with otherlighter species, Aquarius paludum seems stronger, but the ratio of force to weight is actually lower. A. paludum applies about0.3 mN.cm~' to 0.4 mN-cm~' with its mid-legs, thus avoiding penetrating the surface tension layer while propelling itself rapidlyover the water surface. We also investigated the extemal morphology with SEM. The body is covered by efectively two layersof macro and micro-hairs, which renders them hydrophobic. The setae are long (40 μm 60 um) and stif, being responsible forwaterprofing, and the microtrichia are much smaller (<10 um), slender, and flexible, holding a bubble over the body whensubmerged.Keywords: force, surface tension, aquatic insect, morphology, microtrichia, setaeCopyright 。2008, jili University. Published by Science Press and Elsevier Limited. All rights reserved.against the surface waves. Denny'+ pointed out that very1 Introductionsmall insects could not produce waves. Suter et al!5)The insects of the order Hemiptera-Heteroptera,resolved this paradox by showing that in water spidersalso called“bugs" or“true bugs", are small (few milli-the resistance-producing mechanism during the legmeter long) to very big (about 120 mm) animals, whichthrust is the drag. Hu et al,"ol demonstrated this in waterlive virtually in all environments, including water. Thestriders (Aquarius remigis) by analysing the vortices.Gerridae family is one of the largest semiaquatic heter-produced in water by the legs. Bush and HulI suggested'opteran families living in almost all aquatic environ-that curvature forces (i.e. surface tension) would also bements. They are present in big lakes and rivers and smallimportant at least during the rowing gate.ponds as tiny as a tree hole. They are also present in theThe stroke force of water striders was estimated bysea, both at the coast and in the open ocean, being theAndersen21 and later by Hu et al."l. Both of these indi-only insect able to develop the whole life cycle in such arect estimations were based upon the mass and the ac-biotope. These insects have hydrofuge hairs all overceleration of the animal. Perez Goodwyn and Fujisakil8]their body, including their legs". The mid- andmeasured the forces produced by seven different specieshind-legs are the means of propulsion, while the forelegsof water striders directly, but they used a fixed-tetheredare raptorial and short".setup, consisting of a one-dimensional sensor attachedAndersen!2] and Damhofer-Demar')l stated thatdirectlv tp the hack of the animals This prevented themwater striders propel themselves by pushing their legsfrom中国煤化工cing the animals toTHCNMHGCorresponding author; P. J. Perez GoodwynE-mail: pablogoodwyn@yahoo.com.ar122Joumal of Bionic Engineering (2008) Vol.5 No.2push the water back and downwards, increasing the dragz3Dand thus the measured force.∠e Aorce_xsensorWe used a complementary approach involvingin72morphological and biomechanical considerations tofaddress the roles of surface design and force exerting onwater. In this work, our goal was to measure directly and1,accurately the force exerted by a free-swimming water为sstrider during the stroke on water, using a 3D forcesensor.2 Material and method2.1 Force measurementsWater surface-. .. Specimen I VecsoranWe used A. paludum, taken from the Kyoto Uni-Fig.1 Set-up used for the force measurement experiments,versity Insect Ecology Laboratory culture. Five malesshowing the forces' axes and angles considered (see explana-tion in text).(0.03 g 0.032 g) and four females (0.05 g 0.056 g) (allmacropterous) were used for the tests. The experiments1.6were carried out in the IBSS at Nanjing University of.2 tForwardLeftAeronautics and Astronautics. The laboratory tempera-0.8 tture was 22 °C to 24。C, and R.H. was 40%.囝A 60 mm long human hair was attached to thepronotum of the insect using melted beeswax. This was.0 EXaligned with the body axis, and directed posteriorly. The0.4 t甲other end of the hair was glued to a 0.9 mm diameter-0.8 tsteel pin, which was attached to a custom strain gaugeload cell 3D sensor designed at the IBSS (technical de-.180.36 0.540.72tails in Ref. [9]). This force sensor was placed upsideFig.2 Typical force sensor data output, for x, Y and z axesdown approximately 25 mm above the water surface in ashowing three different situations. The first stroke was directed180 mm x 220 mm aquarium in which the water striderright from the sensor, thus the component in X is negative; thesecond is 90" to the X axis (along Y axis), thus the component incould swim. This set-up gave the animal freedom tomove naturally. We defined a 3D Cartesian coordinateis directed to the lef, thus the component in X is positive. Insetsshow diagrammatic high-speed video frames of the corTe-system as shown in Fig. 1. A high-speed video camerasponding strokes.(Mikrotron MC1311, Germany, at 100 fps) was installedalmost perpendicularly (5° to 7" deviated from the per-in the image, and the Z axis was perpendicular to thependicular line over the setup X axis to allow observa-camera image (Fig. 1). The hair glued to the specimention of the sensor, see Fig.1) over the set-up. From thewas positioned at 90° to X axis (i.e. along Y axis). Weimages, we measured the angle a (Figs. 1 and 2) of theselected only those strokes which met two criteria: 1) thehair during the stroke with the software SigmaScan 5.0direction to X axis is within 90*+30*, to minimize the(SPSS) and the angle }B was calculated aseffects of hair bending and to observe the specimenclearly in the high speed video recording; 2) the strokeβ = arsin(,/I),1)was complete at the moment the hair was completelywhere I is the length of the hair and l is projection ofl instretched, transferring the force to the sensor (Fig. 2).the Z direction. The X axis of the sensor was set as theMea中国煤化工was finished and thetop of the camera image, with positive direction to theanimiYHCN M H G possibly underesti-right, the Y axis corresponded to up and down (positive)mating the force), or the stroke had not yet beenPerez Goodwyn et al: Water triders: The Biomechanics of Water Locomotion and FunctionalMorphology of the Hydrophobic Surface (Insecta: Hemiptera-Heteroptera)123completed (possibly overestimating the force), were not3 Resultsconsidered. Thus we chose only the measurements in3.1 Stroke forcewhich the true maximum force was recorded. This re-Data of the resultant force f and the force/weightduced the number of selected strokes to 58 (20 for fe-ratio were pooled for each sex after checking for nor-males and 38 for males) from the approximate 400 re-mality (Shapiro-Wilk W Test P> 0.05). Considering thecorded. The force sensor data was acquired with Lab-force f, the females seem significantly stronger thanview 6.0 (National Instruments, USA) software andthe males (t-test P < 0.0001) (Fig. 3). However theexported to a calculation table sofware program forforce/weight ratios of males and females show no sig-analysis. Knowing the force components,fofs andfz in X,nificant difference between the two sexes (-test P= 0.37)Y and Z axes, the resultant force can be calculated as(Fig. 4).f=2+,+7. (2)OMaleThe ratio of force f (in mN) to weight of the insect1.6■ Female(massxg) (following Perez Goodwyn and Fujisaki8)was defined as an index, a dimensionless performance.The projected angles () offon X, Y and Z axes0一(refer to Fig. 1) are given by0.8 ty = arccos(f,1f)0.6-%2 = arccos(,1f).(3)0.4Y = arccos(f,1{)Applying a purely geometrical approach, the projectedangles y' (see Fig. 1 for details) are given byFig.3 Average (+SD) of the stroke forces measured forAquarius paludum males and females (* indicates signifcantx = arccos(cosβ * cosa)difference t-test P<0.0001).= arccos(osβ* sina) .(4) .。「口Male1=90-β■FemaleThese two methods should yield the same angles in anG. lac1/A. rem.ideal situation. The difference between the angles cal-culated by the two methods was considered to corre-n.s.spond to the error in our calculations, and expressed inpercentage.2Statistical analysis was performed using Statistica6.0 (Stat Soft).2.2 External morphology observationWe observed A. paludum, taken from above men-Fig. 4 Left two columns are average (+ SD) of force/weightratios of Aquarius paludum males and females (n.s. indicatestioned culture. Two males and two females were pre-no significant difference t-test P>0.05). The third column is theserved in 70% alcohol. The samples were allowed to dryforce/weight ratio of A. remigis and Gerris lacustris (based onliterature data) for comparison (no t-test was conducted).at room temperature, and observed in a Keyence VE3000 (Keyence Corp., Osaka, Japan) Scanning Electrone. the error in ourMicroscope without coating at I kV to 3 kV accelerationmeasur中国煤化工inY and 13.7% involtages.z, respeYHCNMHG124Journal of Bionic Engineering (2008) Vol.5 No.2Linear regression shows that the percentage error ofIn contrast, the legs are covered exclusively by se-γ and y, is not significantly correlated with the measuredtae. No microtrichia are visible on the relatively smoothstroke force (error X: R2 = 0.13, P= 0.048; error Y: R2=cuticle. The setae are about 20 μm to 80 μm long, and0.0083, P= 0.49; error z: R2 = 0.0068, P= 0.5378).distally incline. However, their density is higher (27 000setae per mm2) (Fig.5c).3.2 External morphologyThe body cover (Fig. 5a) consists of two different4 Discussionhair layers from different scales and origins, the macro-We successfully developed a method to measureand micro-hair layers. The macro-hair layer is composeddirectly the force exerted by a water strider with a 3Dof long, evenly tapered setae inserted in sockets on theforce sensor. Until now, force has been estimated onlycuticle body surface (Fig. 5b). They are 40 um to 60 μmindirectly either based upon the mass and the accelera-long, about 2 μm wide at the base and incline about 30°tion of the anima14.0, or the momentum imparted in theto 50° relative to the insect surface. The density of setaevortex produced on the water during the legs strokelo.ranges is about 12 000 setae per mm4. Setae are stri-Perez Goodwyn and Fujisaki8 measured directlyated longitudinally with a certain pitch. The grooves arethe force produced by water striders. But they used aca. 200 nm to 400 nm wide.fixed setup, preventing the free movement of the animal.The micro-scale layer consists of 1 μm to 10 μmThis set-up increased dramatically the drag and thus thelong, about 500 nm wide filiform microtrichia, arisingforce measured (4 mN to 6 mN for A. paludum). In theperpendicularly from the cuticle, but irregularly bendingpresent research the force we measured for A. paludumat the apex (Fig. 5b). They occur at a density of 80 000 (Fig. 4) is about one quarter of the force mceasured byto 90 000 microtrichia per mm^. Due to the high aspectPerez Goodwyn and Fujisaki8].ratio, they tend to adhere and collapse in scanning elec-The force measured in this work is higher than thattron microscopy (SEM) preparations!'0.of analytical estimation by other authors2.0. Both theseauthors used different species (Gerris lacustris and A.remigis respectively), achieving the same result, 0.5 mN,but, coincidentally, both species have a mass of about0.01 g. The mass of species we chose (A. paludum) isthree to five times of them. According to Perez Good-wyn and Fujisakis, bigger species are indeed strongerbut their force to weight ratio is lower. Examining thisratio calculated from the data of Andersen'4 and Hu et150 μmal!6l, a comparison can be made (Fig. 4). This situationfits well with our results. The two species with a mass of0.01 g have a much higher force to weight ratio than A.paludum, even though their stroke force is lower.4.1 Error estimationWe are able to estimate the error by comparing thegeometrical data with that derived from the measuredforce and this varies between 5% and 18%. We findclearly different error margins in different axes and thereFig, 5 Water strider Aquarius paludum extemal morphologyare中国煤化工.under SEM. (A) General view, ventral side. (B) Detail of body,|YHC N M H Gsensor was workingshowing setae with underlying cover of thin, flexible micro-trichia. (C) Detail of leg, showing only stiff setae.close to the lower resoluton limt. if this were a cause ofPerez Goodwyn et al: Water Striders: The Biomechanics of Water Locomotion and FunctionalMorphology of the Hydrophobic Surface (Insecta: Hemiptera-Heteroptera)125error, we would expect the lowest force strokes meas-situation is also likely to explain the difference betweenured to have the largest errors, but, as shown in Section 3,errors in different axes.error was not correlated with the force applied. Thus weThe total error is only about 12%, so we can assumecan say that this was not a problem for our method.our data is accurate enough to address the question ofCamera: Another source of error might have beenforce measurement of water striders.the camera position, which was not perfectly perpen-dicular to the set-up but a few degrees tilted (5* to 7"). If4.2 Force and surface tensionthis were true, when the stroke was 90 ° to the X axis, noHu et al!6l suggested that the maximum force perbias in the measurement would have been possible;unit leg length applicable before breaking the free waterhowever, the error in these strokes was not lower. Thus, surface is 1.4 mN~cm '. Perez Goodwyn and Fujisaki8swe can assume that any error caused by this must havefound that A. paludum surpassed this limit, and indeedbeen very low.broke the surface tension layer. But this was due, asHair bending: Human hair has a Young modulus ofexplained in that work, to the fixed set-up.about 5 GPa to 9 GPal, with a diameter of 60 μm to 90We found in this work that A. paludum applied aμm, thus the energy lost due to the bending stiffness mayforce of 0.87 mN to 1.32 mN per stroke. Following thebe very low. The angle measured for our calculationsdeduction of Perez Goodwyn and Fujisakis, this forcemay have changed according to the bending of the hair,was divided by two (left and right legs), from which 70%the force's vector will make a perfect line, but neither thecorresponds to the mid-leg. Thus, according to ourimage (a) nor the trigonometric deduction (B) will be themeasurements, each mid-leg applied 0.3 mN to 0.4 mN,same if the hair is even slightly bent, what actuallyEach mid-leg has approximately 10 mm in contact withhappened. The bending may also explain why the errorthe water (tarsus plus 60% of tibia), Hence, A. paludumis much smaller in the X axis, possibly the hair's weightwould apply 0.3 mN.cm^ “to 0.4 mN.cm，with amight add to the error in the case of both the Y and Zmaximum of 0.6 mN.cm , during the stroke. This givesaxes.a larger safety margin to A. paludum than to A. remigisVector "conversion": The vector of the moving(Hu et al!o, suggested 0.8 mN.cm~) in a dynamicspecimen, should have a large Y or X component, and asituation. However Hu et al.l considered only the midsmall positive Z component, just enough to raise theeg as responsible for 100% of the force and they as-body a few mm above the water surface.sumed that only the tarsi touched with water during theHowever, the force acting on the sensor through thestroke. This mistaken estimation probably means thathair would be slightly different. The component in Zsmaller species, like A. remigis, and G. lacustris havewould be negative and bigger, because the animal wouldactually a greater safety margin than bigger ones. Thisbe pulling the sensor downwards. If we assume that lttlewould in turm support the observation of Hu et al.!ol thatenergy is lost in this force component conversion, the the bigger the species is, the longer the legs are, whichresultant force should be approximately the same, al-avoids breaking the surface tension layer; however, thethough some energy might be lost as hair bending,biggest animals would be close to the size limit for waterstretching and weight lifting, thus creating some error.walkers.System resonance: The signal input can be consid-ered to cause an impact when the specimen stretches the4.3 External morphologyhair and this might lead to a disturbance after the impact,A dense layer of hairs, conferring either a silverybecause of the extra weight added to the sensor, forsheen or matte aspect to the insect surface, covers theinstance the pin and holder. The increased mass in thebodyThis cover differs ac-sensor naturally reduces the system frequency and thuscordinYH中国煤化工n to the water. Thean impact can induce an error in the measurement. ThismainC N M H Gs water protection126Joumal of Bionic Engineeing (2008) Vol.5 No.2making the body efectively waterproof by promotingboratoire de Physique du Solide, Facultes Universitairesfast runoff of water droplets.Notre- Dame de la Paix, Belgium) for checking andIn adult Gerridae, there is a trade off between setae,improving the manuscript.which promote water runoff (waterproofing), and theReferencesmicrotrichia, which maintain a compressible air bubble1] Perez Goodwyn P J. Anti-wetting surfaces in Heteropterain case of submersionl!,". Microtrichia have a very high(Insecta): Hairy solutions to any problem. In: Gorb s N (ed.),aspect ratio, being flexible and with a very high density.Functional Surfaces in Biology, Springer Verlag, Dordrecht,This allows them to keep an air layer over the body if the2007.whole animal is submersed. On the other hand, for wa-2] Andersen N M. A comparative study of locomotion on theterproofing, large spaces are necessary between thewater surface in semiaquatic bugs (Insecta, Hemiptera,structures to decrease the solid to water ratio and in-Gerromorpha). Vidensk Meddr Dansk Naturh Foren, 1976,crease the air to water ratio at the interface (fl and f2,139, 337 -396.respectively, in the Cassie- Baxter lawll). The struc-[3] Darmnhofer-Demar B. On the forward locomotion of thetures must be also stiff enough to resist the capillarywater strider Gerris lacustris L. over the water surface. Zo-attraction of a single droplet over the hair cover. Thus,ologischer Anzeiger Supplemen, 1969, 32, 430 439. (inthe requirements for waterproofing and resisting waterGerman)under pressure are in opposition. Simple waterproofing4] Denny M W. Air and Water. The Biology and Physics ofand water resistance under water pressure are not com-Life's Media, Princeton University Press, Princeton, 1993.pletely compatible. Water striders have found an opti-5] Suter R B, Rosenberg 0, Loeb s, Wildman H, LongJ H.Locomotion on the water surface: propulsive mechanisms ofmum solution, by which in the body both setae and mi-the fisher spider Dolomedes Triton. Journal of Experimentalcrotrichia are present, keeping the body dry whenBiology, 1997, 200, 2523 -2538.splashed, and in case of submersion, holding an air6] Hu D L, Chan B, Bush J W M. The hydrodynamics of waterbubble for a period of time.strider locomotion. Nature, 2003,. 424, 663- 666.65 Conclusion[7] Bush J W M, Hu D L. Walking on water: Biolocomotion atthe iterface. Anual Reriew of Fluid Mechanicas, 2006, 38,We developed a method to measure directly the339- -369.force applied by a water strider on the water surface. Our8] Perez Goodwyn P, Fujisaki K. Sexual conficts, loss of flight,data confirmed that heavier water striders are stronger,and ftness gains in locomotion of polymorphic water strid-while their force/weight ratio is lower. Water stridersers (Gerridae). Entomologia Experimentalis et Applicata,apply enough force to dart over the water surface,2007, 124, 249 -259.without compromising the integrity of the surface ten-9] Zhang ZJ, Ji A H, Wang Z Y, Dai Z D. Three dimensionalsion layer and, in fact, operate within a generous safetysensor for measuring gecko's ground reaction forces. Chi-margin.nese. Journal of Sensors and Actuators, 2007, 20, 1271-1274.Water striders are covered by a double functional(in Chinese)layer of hairs, which makes them both hydrophobic and[10] Andersen N M. Fine structure of the body hair layers andable to submerge for a period of time.morphology of the spiracles of semiaquatic bugs (Insecta,Hemiptera, Geromorpha) in relation to life on the waterAcknowledgementsurface. Vidensk Meddr Dansk Naturh Foren, 1977, 140,We would like to express our sincere thanks to A.7- 37.Matsueda (Kyoto University, Lab. Insect Ecology, Japan)[11] Wortmann F J, Schwan-Jonczyk A. Investigating hair prop-erties relevant for hair “handle". Part I: Hair diameter,for providing specimens necessary for this research. Thefirst author was invited to the IBSS as guest scientist中国煤化Ieatinal Joumal offunded by National Science Foundation of China (Key[12]THc N M H Gry of porouss urfe.sProject No. 60535020). We thank Victoria Welch (La-Transacions of the Faraday Sociery, 1944, 40, 546 -1.